Complexity in the networks of interactions among and between the living and abiotic components forming ecosystems confounds the ability of ecologists to predict the economic consequences of perturbations such as species deletions in nature. Such uncertainty hampers prudent decision making about where and when to invest most intensively in species conservation programmes. Demystifying ecosystem responses to biodiversity alterations may be best achieved through the study of the interactions allowing biotic communities to compensate internally for population changes in terms of contributing to ecosystem function, or their intrinsic functional redundancy. Because individual organisms are the biologically discrete working components of ecosystems and because environmental changes are perceived at the scale of the individual, a mechanistic understanding of functional redundancy will hinge upon understanding how individuals' behaviours influence population dynamics in the complex community setting. Here, I use analytical and graphical modelling to construct a conceptual framework for predicting the conditions under which varying degrees of interspecific functional redundancy can be found in dynamic ecosystems. The framework is founded on principles related to food web successional theory, which provides some evolutionary insights for mechanistically linking functional roles of discrete, interacting organisms with the dynamics of ecosystems because energy is the currency both for ecological fitness and for food web commerce. Net productivity is considered the most contextually relevant ecosystem process variable because of its socioeconomic significance and because it ultimately subsumes all biological processes and interactions. Redundancy relative to productivity is suggested to manifest most directly as compensatory niche shifts among adaptive foragers in exploitation ecosystems, facilitating coexistence and enhancing ecosystem recovery after disturbances which alter species' relative abundances, such as extinctions. The framework further explicates how resource scarcity and environmental stochasticity may constitute ‘ecosystem legacies’ influencing the emergence of redundancy by shaping the background conditions for foraging behaviour evolution and, consequently, the prevalence of compensatory interactions. Because it generates experimentally testable predictions for a priori hypothesis testing about when and where varying degrees of functional redundancy are likely to be found in food webs, the framework may be useful for advancing toward the reliable knowledge of biodiversity and ecosystem function relations necessary for prudent prioritization of conservation programmes. The theory presented here introduces explanation of how increasing diversity can have a negative influence on ecosystem sustainability by altering the environment for biotic interactions – and thereby changing functional compensability among biota – under particular conditions.